A rapid detection method and system for iron ions in environmental water

Abstract

The present invention belongs to the technical field of analysis and detection, and specifically relates to a method and system for rapid detection of iron ions in environmental water. The method of the present invention comprises the following steps: step 1, treating a water sample to be detected with a chemical color development system, causing the chemical color development system to change color, and taking a picture of the chemical color development system; step 2, segmenting a region of interest from the picture; step 3, calculating the change value of each component in the RGB color space and the HSI color space in the region of interest; step 4, using the change value of at least two components obtained in step 3 as a feature, analyzing the content of iron ions in the water sample to be detected by a multivariate regression equation model. The present invention has significant advantages such as low cost, simple operation, high detection efficiency, and strong anti-interference ability, and is suitable for detecting Fe ions in natural water bodies. 3+ The rapid detection of concentration has a good application prospect.

Description

Technical Field

[0001] The present invention belongs to the technical field of analysis and detection, and in particular relates to a method and system for rapid detection of iron ions in environmental water. Background Art

[0002] Iron (Fe) is a critical element in the ecological environment and an essential trace element for the human body. In natural waters, Fe plays a crucial role in the transfer of oxidative stress, and its valence and content play a crucial role in maintaining aquatic quality. Therefore, information on Fe concentrations in natural waters plays a crucial role in fields such as environmental science, environmental toxicology, life sciences, and food safety.

[0003] Existing analytical methods for detecting Fe in water primarily include spectroscopy and mass spectrometry. While these methods offer high precision and sensitivity, they suffer from common limitations, such as bulky and expensive analytical equipment and demanding experimental conditions, making them difficult to apply to field-based Fe detection. Therefore, developing a method for analyzing Fe in natural water that is highly accurate, low-cost, simple to operate, and rapid is of great significance.

[0004] Chemical colorimetric detection of Fe is a low-cost, simple, and rapid method for determining Fe in water (Xu Xiawei et al., Determination of Iron in Water by Iron Colorimetric Detection Plate). The principle is to use hydroxylamine hydrochloride to reduce the trivalent iron ions in water to divalent iron, and then use the colorimetric reaction results of divalent iron and 1,10-phenanthroline to detect the iron element. However, since the colorimetric reaction reflects the detection results through color changes, if the colorimetric results are to be accurately and quantitatively analyzed, equipment such as spectrometers are usually required. This has a very adverse effect on the cost and ease of application of the method for chemical colorimetric detection of Fe elements. Therefore, how to accurately analyze the results of chemical colorimetric development using a simpler and faster method remains a problem that urgently needs to be solved in this field. Summary of the Invention

[0005] In response to the problems of the prior art, the present invention provides a method and system for rapid detection of iron ions in environmental water, the purpose of which is to use a multivariate regression equation model to analyze images of a chemical color development system to achieve high-precision, low-cost, simple and rapid detection of iron content.

[0006] A rapid detection method for iron ions in environmental water comprises the following steps:

[0007] Step 1: treating the water sample to be tested with a chemical color development system to cause the chemical color development system to change color, and taking a picture of the chemical color development system;

[0008] Step 2: Segmenting a region of interest from the image;

[0009] Step 3, calculating the change value of each component of the RGB color space and the HSI color space in the region of interest;

[0010] Step 4: Using the change values ​​of at least two components obtained in step 3 as features, the content of iron ions in the water sample to be tested is analyzed through a multivariate regression equation model.

[0011] Preferably, in step 1, the chemical color development system includes hydroxylamine hydrochloride and 1,10-phenanthroline.

[0012] Preferably, in step 1, the chemical color development system is attached to the microfluidic paper chip, and the amount of hydroxylamine hydrochloride attached is 0.622-1.243 mg / cm 2 The adhesion amount of the 1,10-phenanthroline is 0.100-0.199 mg / cm 2 , the color development time is 15-25min.

[0013] Preferably, in step 2, an edge segmentation method is used to segment the region of interest.

[0014] Preferably, in step 2, the image is subjected to median filtering to segment the image into regions of interest.

[0015] Preferably, in step 3, the change value of the component is calculated according to the following formula:

[0016]

[0017] Where M represents the component of RGB color space or HSI color space, which can be one of R, G, B, H, I or S; M i is the value of M in group i among n parallel experimental groups; M n+1 It is the blank control value of the chemical colorimetric system M with blank water sample added.

[0018] Preferably, in step 4, the change value of the component selected as the feature is and in is the change value of R, is the change value of B, is the change value of S.

[0019] Preferably, in step 4, the multivariate regression equation model is:

[0020]

[0021] The present invention also provides a rapid detection system for iron ions in environmental water, comprising:

[0022] A camera module for taking pictures of the chemical color development system;

[0023] a segmentation module, for segmenting a region of interest from the image of the chemical color development system;

[0024] The calculation module is used to calculate the change value of each component of the RGB color space and the HSI color space in the region of interest; and use the change values ​​of at least two components as features to analyze the content of iron ions in the water sample to be tested through a multivariate regression equation model.

[0025] The present invention also provides a computer-readable storage medium storing a computer program for implementing the above-mentioned method for rapid detection of iron ions in environmental water.

[0026] The present invention reacts the water sample to be tested with the chemical color development system, takes a picture of the chemical color development system with an industrial camera, and then uses a multivariate regression equation model to analyze the color changes in the picture, thereby realizing the analysis of the iron content in the water sample to be tested. In a preferred embodiment, the present invention optimizes the dosage of the reducing agent and the color developer in the chemical color development system, the method of segmenting the region of interest, and the multivariate regression equation model, further improving the accuracy of the analysis results. In a preferred embodiment of the present invention, the determination coefficient R of the test result is 2 The detection limit is 0.9932, the mean square error MSE is 5.6491, the root mean square error RMSE is 2.3768, the detection limit is 2.53μM, the detection range is 5-100μM, the relative standard deviation RSD of 10 parallel measurements is less than 3.0%, the spike recovery rate is 94%, and the cost of a single detection is about 0.7 yuan. In addition, the method and system of the present invention are simple to operate and efficient. It is only necessary to perform a color reaction on the water sample to be tested, take a picture and input the picture into the computer to obtain the test result. The detection cycle is about 18min, the operation is simple, and the results are obtained quickly. It can be seen that the present invention has significant advantages such as low cost, simple operation, high detection efficiency, and strong anti-interference ability, and is suitable for Fe in natural water bodies. 2+ and Fe 3+ The rapid detection of concentration has a good application prospect.

[0027] Obviously, based on the above contents of the present invention, according to common technical knowledge and customary means in this field, without departing from the above basic technical ideas of the present invention, other various forms of modifications, replacements or changes can be made.

[0028] The following further describes the above content of the present invention in detail through specific embodiments in the form of examples. However, this should not be construed as limiting the scope of the above subject matter of the present invention to the following examples. All technologies implemented based on the above content of the present invention fall within the scope of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 Design diagram for μPADs.

[0030] Figure 2 The μPADs images before and after adding noise and the filtering effect diagram, Fe 3+ The concentration is 100 μM. (a) Original image before adding noise; (b) Image after adding noise; (c) Image after mean filtering of the image after adding noise; (d) Image after median filtering of the image after adding noise.

[0031] Figure 3 The histograms of the μPADs image before and after adding noise and the filtering effect. (a) Histogram before adding noise; (b) Histogram after adding noise; (c) Histogram after mean filtering; (d) Histogram after median filtering.

[0032] Figure 4 The original grayscale image of the μPADs detection area and the effects of different segmentation methods. The Fe 3+ The concentration is 100 μM. (a) Original grayscale image of the μPADs detection area; (b) Threshold segmentation method; (c) Edge segmentation method; (d) Region growing method.

[0033] Figure 5 This is a graph showing the relationship between hydroxylamine hydrochloride concentration and color difference d.

[0034] Figure 6 This is a graph showing the relationship between 1,10-phenanthroline concentration and color difference d.

[0035] Figure 7 For different concentrations of Fe 3+ μPADs image of the sample solution (after segmentation).

[0036] Figure 8 A single eigenvalue with Fe 3+ The relationship between the concentration of the solution. (a) with Fe 3+ Relationship diagram between solution concentration; (b) with Fe 3+ Graph showing the relationship between solution concentrations.

[0037] Figure 9 A single eigenvalue with Fe 3+ The relationship between the concentration of the solution. (a) with Fe 3+ Relationship diagram between solution concentration; (b) with Fe 3+ Graph showing the relationship between solution concentrations.

[0038] Figure 10 A single eigenvalue with Fe 3+ The relationship between the concentration of the solution. (a) with Fe 3+ Relationship diagram between solution concentration; (b) with Fe 3+ Graph showing the relationship between solution concentrations.

[0039] Figure 11 For different interfering ions 100.0 μM Fe 3+ The dotted line in the figure represents the concentration of the control group. DETAILED DESCRIPTION

[0040] It should be noted that the algorithms for data collection, transmission, storage and processing steps not specifically described in the embodiments, as well as the hardware structures, circuit connections, etc. not specifically described can all be implemented through the disclosed content of the prior art.

[0041] Example 1 Iron ion detection method and system based on multivariate regression equation model

[0042] This embodiment provides an iron ion detection system based on a multivariate regression equation model, comprising:

[0043] A camera module for taking pictures of the chemical color development system;

[0044] a segmentation module, for segmenting a region of interest from the image of the chemical color development system;

[0045] The calculation module is used to calculate the change value of each component of the RGB color space and the HSI color space in the region of interest; and use the change values ​​of at least two components as features to analyze the content of iron ions in the water sample to be tested through a multivariate regression equation model.

[0046] The chemical color development system includes hydroxylamine hydrochloride and 1,10-phenanthroline attached to the microfluidic paper chip, and the attachment amount of the hydroxylamine hydrochloride is 0.622 mg / cm 2 The adhesion amount of 1,10-phenanthroline is 0.100 mg / cm 2 .

[0047] The design of the microfluidic paper chip (μPADs) is as follows:

[0048] The specific structure of μPADs is as follows Figure 1The μPAD substrate is a 5 cm × 5 cm hydrophobic paper (see the gray area in the figure). Four circular hydrophilic paper discs with a diameter of 16 mm (see the white area in the figure) are attached to the hydrophobic paper. The centers of the four circular discs form a square with a side length of 2 cm. Discs 1-3 are the test area, used for three replicate measurements, and disc 4 is used as a blank control.

[0049] The method for detecting the iron content using the above system includes the following steps:

[0050] Step 1: Treat the water sample to be tested with a chemical color development system to cause the chemical color development system to change color, and then take a picture of the chemical color development system. The specific steps are as follows:

[0051] (1) First, take a blank μPAD. Then, add 25 μL of hydroxylamine hydrochloride / 1,10-phenanthroline mixed solution to each of the four circular paper discs and let it stand for 25 min.

[0052] (2) When the μPADs were completely dry, 15 μL of 1.4 g / L polyacrylic acid solution was added to each of the four circular paper pieces and allowed to stand for 25 min. When the μPADs were completely dry, the pretreatment process of the μPADs was completed.

[0053] The pretreated μPADs can be directly used to detect Fe in water. 3+ Concentration. The specific detection steps are as follows:

[0054] (1) Take a pre-treated μPAD. Then, add 25 μL of the sample to be tested in the detection areas 1-3 of the μPAD. Then, add 25 μL of dilute hydrochloric acid solution with a pH of 1.5 to the blank control area 4.

[0055] (2) After the μPADs were allowed to stand for 15 minutes, an industrial camera was used to capture images of the μPADs after chemical color development. Figure 1-3 The specific process is as follows: Place the chemically developed μPADs on a stage with a distance of 10 cm between the stage and the industrial camera lens, and position the ring light source at the same height as the lens surface. Next, adjust the camera aperture and light source brightness to achieve optimal image brightness, maintaining consistent shooting conditions. Finally, use VisionBank MV machine vision measurement software on a computer to control the industrial camera and capture images of the chemically developed μPADs. The images are stored in TIFF format with a size of 1944 × 2592.

[0056] Step 2: Perform median filtering on the image to segment the region of interest from the image using an edge segmentation method.

[0057] Step 3, calculating the change value of each component of the RGB color space and the HSI color space in the region of interest;

[0058] The change value of the component is calculated according to the following formula:

[0059]

[0060] Where M represents the component of RGB color space or HSI color space, which can be one of R, G, B, H, I or S; M i is the value of M in group i among n parallel experimental groups; M n+1 is the blank control value of the chemical color development system M with blank water sample added. In this embodiment, three circular paper discs for parallel experimental groups are set on the μPADs, so n=3.

[0061] Step 4: Using the change values ​​of at least two components obtained in step 3 as features, the content of iron ions in the water sample to be tested is analyzed through a multivariate regression equation model.

[0062] As a preferred solution, the multivariate regression equation model in this embodiment is:

[0063]

[0064] Wherein, Y is the content of iron, in μM. is the change value of R, is the change value of B, is the change value of S.

[0065] The following experiments further illustrate the technical solution of the present invention. The steps not specifically described in the following experimental examples are the same as those in Example 1.

[0066] Experimental Example 1 Filtering Method Screening

[0067] This experimental example screens the filtering method in step 2.

[0068] Images developed using μPADs using industrial cameras often contain noise. This noise, primarily due to illumination and the camera itself, appears randomly as isolated points. This experiment uses MATLAB software to perform mean and median filtering on images developed using μPADs, and then conducts comparative analysis of the resulting images. The experimental steps are as follows:

[0069] (1) First, use the imread function in MATLB software to read the color image of μPADs.

[0070] (2) Then, uniformly randomly distributed noise is added to the image. The original image and the image after adding noise are shown in Figure 2. Figure 2 (a) and Figure 2 (b).

[0071] (3) After that, the color space of the image is converted from RGB to HSI, and then the three sub-channel images of HSI are filtered separately.

[0072] (4) Convert the filtered channel-wise image from the HSI color space to the RGB color space.

[0073] Mean filtering. Use the fspecial function in MATLAB to generate a 2D digital image filter. The type and parameters are set to average and 3 respectively. Then use the filter2 function to perform mean filtering on the image after adding noise. The filter window is a 3×3 square window. The filtered image is shown in Figure 2 (c).

[0074] Median filter processing. The medfilt2 median filter in MATLAB is used to perform median filtering on the image after adding noise. The parameters are set to [3,3] and the filter window is a 3×3 square window. The filtered image is shown in Figure 2 (d).

[0075] When using mean filter to smooth the preprocessed image, the image becomes more blurred as the window expands. This is most obvious at the edge of the image. The median filter has a relatively good noise reduction capability and retains more details of the image. Figure 3 It can be seen that the image histogram after median filtering (see Figure 3 (d)) is closest to the original image. Therefore, median filtering is a better choice for smoothing the original data image.

[0076] Experimental Example 2: Screening of Region of Interest Segmentation Methods

[0077] The original data image (photographed image) acquired by this method contains the μPADs and the background, while the color feature data is concentrated in four circular areas of the μPADs. To accurately obtain the image data of these four areas, the original image needs to be segmented and the region of interest (ROI) within the image needs to be extracted.

[0078] To this end, this experiment investigated the effectiveness of threshold segmentation, edge segmentation, and region growing in isolating the ROI region from the μPADs image. The images used for segmentation were all partial regions of the μPADs after color development, which contained a circular hydrophilic region (i.e., the ROI region).

[0079] First, use the imread function in the MATLB software to read the image of the μPADs, and then use a 3×3 square window to perform median filtering on the image. Then, use the rgb2gray function to convert the image into a grayscale image, as shown in Figure 4 (a). Finally, Figure 4 (a) Perform segmentation.

[0080] (1) Threshold segmentation method. The grayscale image is segmented by the histogram method using the graythresh function in MATLAB software, and then the grayscale image is binarized using the im2bw function. The results are shown in Figure 4 (b) White represents the circular hydrophilic area, and black represents the wax paper background area. This method is computationally simple, but the segmented image edges may contain many burrs, and the background area may contain isolated white areas, resulting in insufficient recognition between the target and the background.

[0081] (2) Edge segmentation method. According to the image size, after calculation and measurement, it can be known that the radius of the circular area ranges from 256 to 258. Using MATLAB software, the center of the circular contour is detected by Hough transform. According to the obtained radius and center position, a circle is drawn and the segmented image is shown in FIG. Figure 4 (c) The white area in the figure represents the hydrophilic region, and the black area represents the background. This method separates the target region as a complete circle, but the process of identifying the circle's center can sometimes cause deviations, causing the final segmented region to deviate from the true position of the circular region.

[0082] (3) Region growing method. The grayscale threshold is determined using the graythresh function in MATLAB software. After that, a point in the background area is selected as the initial pixel point. The similarity is determined based on the threshold, and the pixels in the region are merged to achieve image segmentation. Figure 4 (d) shows the binary image of the segmentation result, where white represents the circular hydrophilic area and black represents the wax paper background area. This method can eliminate isolated points within the target and background areas, but the edge contours of the separated target image still have burrs.

[0083] After comparing the three segmentation methods, it was found that the edge segmentation method was the better choice, which could ensure that the target area was circular after separation and effectively retain the effective color information in the μPADs detection area.

[0084] Experimental Example 3 Optimization of color development conditions

[0085] 1. Experimental Methods

[0086] 1. Optimal Condition Judgment Method

[0087] In the same color space, color difference represents the degree of difference between two colors and can be measured by the Euclidean distance between the coordinates of the two color parameters. In a two-dimensional digital image, the mean of each color component within a region can be used to generate a new color that represents the average color of the region. The color difference between the average colors of the regions represents the color difference between the regions. For μPAD-based color images, the degree of color rendering can be measured by the color difference d, which is calculated as follows:

[0088]

[0089] in, They represent the differences between the average values ​​of the color components R, G, and B in the μPADs color image detection area and the blank control area, respectively.

[0090] In this method, the larger the color difference d, the more obvious the color rendering of the μPADs, the more effective information in the color image, and the more conducive to the subsequent extraction of color features. Therefore, in the color rendering condition optimization experiment, the color rendering condition corresponding to the maximum d value was selected as the optimal condition.

[0091] 2. Optimization of hydroxylamine hydrochloride dosage

[0092] (1) Preparation of solution

[0093] First, 200 g / L hydroxylamine hydrochloride solution was diluted with 6.3 M acetate buffer (pH = 4.5) to obtain hydroxylamine hydrochloride solutions with concentrations of 100 g / L, 50 g / L, 20 g / L, and 2 g / L, respectively.

[0094] Then, accurately pipette 1 mL of 6.3 M acetate buffer with a pH of 4.5 into 2 mL of 2 g / L hydroxylamine hydrochloride solution. After uniform mixing, add 1 mL of 32 g / L 1,10-phenanthroline solution to the mixed solution. After uniform mixing, the resulting mixed solution is recorded as mixed solution No. 1.

[0095] Afterwards, the above preparation process was repeated using hydroxylamine hydrochloride solutions of another four concentrations to prepare four mixed solutions, which were respectively designated as mixed solutions No. 2 to No. 5.

[0096] The concentrations of hydroxylamine hydrochloride in the experimentally prepared mixed solutions No. 1-5 were 1 g / L, 10 g / L, 25 g / L, 50 g / L, and 100 g / L, respectively, and the concentration of 1,10-phenanthroline was 8 g / L.

[0097] (2) Pretreatment of μPADs

[0098] Prepare five sets of blank μPADs, each containing three blank μPADs (groups 1-5). First, add 25 μL of the hydroxylamine hydrochloride / 1,10-phenanthroline mixture to each of the four hydrophilic circular areas of the μPADs. The mixture numbered for each μPAD group matches the group number, and the experimental procedure for each μPAD group is identical. Allow the μPADs to dry completely for 25 minutes, then add 15 μL of the polyacrylic acid solution to each of the four hydrophilic areas. Finally, allow the μPADs to dry. The resulting μPADs can be used for sample testing in the hydroxylamine hydrochloride concentration optimization experiment.

[0099] (3) Acquisition and processing of μPADs color images

[0100] Take a μPAD treated in step (2) and add 25 μL of 100 μM Fe 3+ Solution, 25μL of pH=1.5 dilute hydrochloric acid solution was added to the control area as a blank control. As the solution gradually diffused under the capillary action, the coloring area began to show orange-red. The closer to the edge of the hydrophilic area, the lighter the orange-red. As the color development reaction proceeded, the orange-red color of the coloring area gradually deepened. After the μPADs were allowed to stand for 15 minutes, the color development image of the μPADs was captured using an industrial camera. Finally, the color development image was processed using MATLAB software and color features were extracted. Calculate the color difference d.

[0101] 3. Optimization of 1,10-phenanthroline dosage

[0102] (1) Preparation of solution

[0103] 32 g / L of 1,10-phenanthroline solution was diluted with 6.3 M acetate buffer (pH=4.5) to obtain 1,10-phenanthroline solutions with concentrations of 16 g / L, 8 g / L, and 1.6 g / L, respectively.

[0104] Accurately pipette 1 mL of 6.3 M acetate buffer at pH 4.5 and add it to 2 mL of 1.6 g / L 1,10-phenanthroline solution. After uniform mixing, add 1 mL of 200 g / L hydroxylamine hydrochloride solution to the mixed solution and mix well. The resulting mixed solution is recorded as mixed solution No. 6.

[0105] The above preparation process was repeated using 1,10-phenanthroline solutions of three other concentrations to prepare three mixed solutions, which were respectively designated as mixed solutions No. 7 to No. 9.

[0106] The concentrations of 1,10-phenanthroline in the experimentally prepared mixed solutions No. 6-5 were 0.08 g / L, 0.8 g / L, 8 g / L, and 16 g / L, respectively, and the concentration of hydroxylamine hydrochloride was 50 g / L.

[0107] (2) Pretreatment of μPADs

[0108] Prepare five sets of blank μPADs, each containing three blank μPADs (groups 6-9). First, add 25 μL of the mixture to each of the four hydrophilic circular areas of the μPADs. The number of the mixture added to each μPAD group matches the group number, and the experimental procedure for each μPAD group is the same. After the μPADs are completely dry, add 15 μL of the polyacrylic acid solution to each of the four hydrophilic areas. Finally, let the μPADs dry. The resulting μPADs can be used to test samples in the 1,10-phenanthroline concentration optimization experiment.

[0109] (3) Acquisition and processing of μPADs color images

[0110] Take a μPAD treated in step (2) and add 25 μL of 100 μM Fe 3+ Solution, 25 μL of pH=1.5 dilute hydrochloric acid solution was added to the control area as a blank control. Then, the μPADs were left to stand for 15 minutes, and then the color image of the μPADs was captured using an industrial camera. Finally, the color image was processed using MATLAB software and color features were extracted. Calculate the color difference d.

[0111] 2. Experimental Results

[0112] 1. Optimization of hydroxylamine hydrochloride dosage

[0113] The results are as follows Figure 5 As shown. The experiment shows that as the concentration of hydroxylamine hydrochloride increases, the color difference d gradually increases. When the concentration of hydroxylamine hydrochloride is 50g / L, the d value reaches the maximum. When the concentration of hydroxylamine hydrochloride exceeds 50g / L, the d value decreases, but the change trend is relatively stable. In summary, when the concentration of hydroxylamine hydrochloride is 50g / L, it is the best dosage. After conversion, when the concentration of hydroxylamine hydrochloride is 50g / L, its dosage on μPADs is 0.622mg / cm 2 .

[0114] 2. Optimization of 1,10-phenanthroline dosage

[0115] The color difference d is used to measure the degree of color difference. The effect of 1,10-phenanthroline concentration on the color difference d is as follows: Figure 6As shown. The results show that as the concentration of 1,10-phenanthroline increases, the color difference d gradually increases. When the concentration of 1,10-phenanthroline exceeds 8g / L, the d value is the largest. When the concentration of 1,10-phenanthroline exceeds 8g / L, the d value becomes smaller, but the change trend is smaller than that at low concentrations. In summary, when the concentration of 1,10-phenanthroline is 8g / L, it is the best dosage. After conversion, when the concentration of 1,10-phenanthroline is 8g / L, its dosage on μPADs is 0.100mg / cm 2 .

[0116] Experimental Example 4: Construction of a multivariate regression equation model

[0117] 1. Acquisition of digital image dataset of iron ion paper chip

[0118] To construct a quantitative detection model, we first need to measure different concentrations of Fe 3+ Standard sample solution, obtain the color characteristic value of ROI in the color image Provide a data basis for the research of quantitative models. The specific steps are as follows:

[0119] (1) First, take 250 μL of 10 mM Fe 3+ The stock solution was diluted with a dilute hydrochloric acid solution of pH = 1.5 and fixed to volume in a 250 mL constant volume flask to obtain a concentration of 100 μM Fe 3+ Then, use the above dilute hydrochloric acid solution to 3+ The working solution was further diluted to obtain a series of Fe concentrations between 0-100 μM. 3+ Test solution, the series Fe 3+ The concentrations of the test solutions were 0μM, 5μM, 10μM, 15μM, 20μM, 25μM, 30μM, 35μM, 40μM, 45μM, 50μM, 55μM, 60μM, 65μM, 70μM, 75μM, 80μM, 85μM, 90μM, 95μM, and 100μM, with a gradient of 5μM, for a total of 21 concentration gradients.

[0120] (2) Then, take a pre-treated μPAD and drop 25 μL of Fe at a predetermined concentration on the detection areas 1-3. 3+ For the test solution, 25 μL of the above dilute hydrochloric acid solution was added to the control area No. 4 as a blank control. After the sample was colored on the paper chip, it was left to stand for 15 minutes, and then the μPADs were placed on the stage to take pictures and obtain a digital image in tiff format.

[0121] (3) Then, using MATLAB software, median filtering and edge segmentation were performed on the image. The RGB color components of the hydrophilic areas 1-4 were extracted and converted into HSI color components. Then, for each disc area, the average value of each color component was taken, and the arithmetic mean of the color component means of the three detection areas was subtracted from the corresponding color component mean of the control area to obtain These 6 average values ​​increased.

[0122] (4) Finally, repeat the above experimental steps (2) and (3) to collect digital images of each concentration gradient sample in the concentration range of 0-100 μM. Ten parallel experiments were carried out for each concentration gradient, and a total of 210 Fe 3+ Digital image of μPADs.

[0123] Different concentrations of Fe 3+ The digital image of μPADs is as follows Figure 7 Apparently, when Fe 3+ When the solution concentration increases, the μPADs detection area shows obvious color changes.

[0124] The collected μPADs chemical color images were further processed using MATLAB software to extract and calculate the The six types of mean increment data obtained in 10 parallel experiments were then arithmetic averaged, and the obtained means represented the mean value of each concentration of Fe 3+ Color characteristic value corresponding to the solution Different concentrations of Fe 3+ The color characteristic values ​​of the μPADs chemical colorimetric digital images of the samples are shown in Table 1.

[0125] Table 1 Different concentrations of Fe 3+ Color characteristic values ​​of μPADs chemical color digital image of the sample

[0126]

[0127] 2. Digital Image Quantification Model of Iron Ion Concentration

[0128] Fe 3+ The μPADs digital image quantitative model of Fe concentration in water samples is a 3+ The key to concentration measurement is that its reliability directly affects the accuracy of quantitative analysis. These six color features were studied in relation to the Fe 3+ The quantitative relationship between the concentration and the amount of the product is established, and the corresponding quantitative model is established.

[0129] First, simple regression and multiple regression were used to construct the detection model, and the determination coefficient R 2 , Mean Square Error (MSE) and Root Mean Square Error (RMSE) are used to evaluate the regression effect. The calculation formula is as follows:

[0130]

[0131]

[0132]

[0133] Among them, y i represents the true value of the concentration of the i-th sample, Y i represents the predicted value of the concentration of the i-th sample, represents the average concentration of all samples, and m represents the total number of samples.

[0134] (1) Univariate regression analysis

[0135] To study the relationship between the eigenvalues ​​and Fe 3+ The linear correlation between the concentrations is analyzed. In this experiment, the regress function in MATLAB software was used to perform a univariate regression analysis on the six color feature values. 3+ The relationship between the concentrations is shown in the figure. Figure 8-10 .

[0136] After calculation, among these single components, with Fe 3+ The linear correlation between the concentrations is obvious, and the first-order linear regression equation is:

[0137]

[0138] Where Y is the Fe content in the solution to be tested 3+ Concentration, unit is μM. Determination coefficient R of the fitting equation 2 The mean square error (MSE) is 7.3386, and the root mean square error (RMSE) is 2.7090.

[0139] (2) Multiple linear regression

[0140] From the above univariate regression analysis, we can see that among the single eigenvalue components, except In addition, the other five eigenvalues ​​are related to Fe 3 +The linear relationship between the concentrations is not obvious. Therefore, this experiment divides the six features into two categories based on the RGB and HSI color spaces, performs multiple linear regression on the components of each of the two color spaces, and analyzes the prediction effect of the fitting equation.

[0141] This paper uses the regress function in MATLAB software to Perform multiple linear regression, and then Perform multiple linear regression, and the regression equation is as follows:

[0142]

[0143]

[0144] Where Y represents the Fe content in the sample to be tested 3+ Concentration, in μM. The coefficient of determination R of the regression equation 2 The mean square error (MSE) is 65.7714, and the root mean square error (RMSE) is 8.1100. The coefficient of determination R of the regression equation 2 The mean square error (MSE) is 5.6916, and the root mean square error (RMSE) is 2.3857.

[0145] Experimental results show that the multivariate linear regression effect of HSI color space is better than the fitting effect of all single components. For multivariate linear regression in a single color space, HSI has a better fitting effect.

[0146] To further explore the relationship between the characteristic value and Fe 3+ The relationship model between the concentration and the color space is studied in this experiment. In order to consider the components of the two color spaces at the same time, this paper selects one component from each of the RGB and HSI color spaces and performs binary linear regression. For the HSI color space, only with Fe 3+ The linear correlation between the concentrations is obvious. Therefore, in the HSI color space, the eigenvalue should be selected first.

[0147] The regression equations obtained by three different feature selection methods are as follows:

[0148]

[0149]

[0150]

[0151] The coefficient of determination R of the regression equation 2 The mean square error (MSE) is 5.7127, and the root mean square error (RMSE) is 2.3901. The coefficient of determination R of the regression equation 2 is 0.9930, the mean square error MSE is 5.7976, and the root mean square error RMSE is 2.4078; The coefficient of determination R of the regression equation 2 The mean square error (MSE) is 5.7295, and the root mean square error (RMSE) is 2.3936. The experimental results show that the goodness of fit of the binary linear regression in the two color spaces is similar.

[0152] Based on the above binary linear regression, this paper selects two components from the RGB color space and combines them with the components in the HSI space. Study the fitting effect of its multiple regression. The regression equations corresponding to different component combinations are as follows:

[0153]

[0154]

[0155]

[0156] The coefficient of determination R of the regression equation 2 is 0.9931, the mean square error MSE is 5.7031, and the root mean square error RMSE is 2.3881; The coefficient of determination R of the regression equation 2 is 0.9931, the mean square error MSE is 5.7044, and the root mean square error RMSE is 2.3884; The coefficient of determination R of the regression equation 2 The mean square error (MSE) is 5.6491, and the root mean square error (RMSE) is 2.3768.

[0157] 3. Selection of concentration calculation model

[0158] The fitting results of the above 9 regression equations are shown in Table 2. The MSE and RMSE of the multiple linear regression equation are the smallest, and R 2 The largest, the best fit.

[0159] Table 2 Performance comparison of different regression methods

[0160]

[0161] Therefore, the color feature component is selected To calculate Fe 3+ Concentration is the better choice.

[0162] Experimental Example 5 Method Analysis Performance

[0163] This experimental example studies the analytical performance of the method of Example 1.

[0164] 1. Anti-interference capability

[0165] The actual natural water sample matrix is ​​complex and contains many interfering ions, which may interfere with the measurement results of this method. Therefore, this experimental example examines the ability of this method to resist interference from other metal ions. The specific experimental steps are as follows:

[0166] (1) First, pipette 20 μL of 1000 mg / L Mg 2+ Solution to 2 mL of 100 μM Fe 3+ After mixing evenly, the resulting solution is recorded as sample No. 1. At this time, Mg 2+ The concentration of interfering ions was 10.0 mg / L.

[0167] (2) Repeat step (1) and use Pb 2+ 、Ni 2+ 、Cu 2+ 、Zn 2+ 、Co 2+ , Ca 2+ The sample solutions were prepared by using these 6 single interfering ion solutions. The obtained sample solutions were recorded as samples 2-7, and the interfering ion concentration in the samples was 10.0 mg / L. Then, 20 μL of ultrapure water was pipetted into 2 mL of 100 μM Fe 3+ After mixing evenly, the resulting solution was recorded as sample No. 8 and used as a control experiment.

[0168] (3) Next, μPADs were used to detect samples 1-8, with sample 8 used as a blank control experiment. The color images were captured using an industrial camera.

[0169] (4) Finally, MATLAB software was used to process the μPADs digital images corresponding to samples 1-8 and extract the image color feature values ​​to calculate Fe 3+ Concentration measurement value.

[0170] The experimental results show that (such as Figure 11 ), when the interfering ion concentration in the sample is 10.0 mg / L, the Fe 3+The concentration measurement value had no significant change, and the relative standard deviation (RSD) was 2.93%. The experimental results showed that under the interference of metal ions at a concentration of 10.0 mg / L, the μPADs proposed in the present invention had good anti-interference ability.

[0171] Increase the amount of interfering ions. When the concentration of interfering ions in the sample solution is 100.0 mg / L, for Mg 2+ , Pb 2+ 、Ni 2+ 、Zn 2+ , Ca 2+ For these five heavy metal ions, μPADs still have good anti-interference ability, but cannot overcome Cu 2+ 、Co 2+ At this time, Co 2+ Reacts with 1,10-phenanthroline to form an orange complex, Cu 2+ will make μPADs appear obvious blue. However, in natural water, Cu 2+ and Co 2+ The concentration level of Cu is generally lower than 10.0 mg / L. Therefore, the method of the present invention is not affected by the presence of Cu in the sample. 2+ and Co 2+ impact.

[0172] In summary, the method proposed in the present invention can meet the requirements of Fe 3+ It can be used for quantitative analysis of metal ions and has good anti-interference ability of metal ions.

[0173] 2. Detection Limit

[0174] According to the International Union of Pure and Applied Chemistry (IUPAC), the limit of detection (LOD) is defined as the concentration corresponding to a signal-to-noise ratio of three times the analytical signal. The detection limit of the method of the present invention was investigated, and the experimental steps are as follows:

[0175] (1) A hydrochloric acid solution with a pH of 1.5 was used as a blank sample. The blank sample was detected 10 times using μPADs, and the color image was captured using an industrial camera.

[0176] (2) Use MATLAB software to process the color image and extract the color feature values ​​of the color image, and calculate the standard deviation of each color feature value in 10 experiments.

[0177] (3) Calculate the Fe content value corresponding to the color characteristic value of 3 times the standard deviation, and this value is the detection limit.

[0178] Through blank sample experiments with μPADs, the detection limit of the proposed method was calculated to be 2.53 μM. The National Standard for Drinking Water Quality (GB 5749-2006) and the Environmental Quality Standard for Surface Water (GB3838-2002) stipulate that the Fe content should not exceed 0.3 mg / L, which translates to a mass concentration of 5.37 μM. Therefore, the detection limit of the proposed method meets the requirements for Fe detection in natural waters.

[0179] 3. Spike recovery

[0180] To examine the accuracy and analytical capabilities of this method, a water sample was collected from the Qingshui River in Chengdu's High-tech West District to conduct a spike recovery experiment. The specific steps are as follows:

[0181] (1) First, 500 mL of water sample was taken from the Qingshui River. The water sample was filtered using a suction filter and a 0.45 μm pore size filter membrane and then stored. Then, 250 μL of concentrated hydrochloric acid solution was transferred to the 250 mL water sample and mixed evenly to obtain an acidified water sample.

[0182] (2) Pipette 100 μL of 10 mM Fe at pH 1.5 3+ The solution was added to 99.90 mL of water sample and mixed evenly to obtain spiked sample No. 1. Then, 200 μL of the above Fe 3+ The solution was added to 99.8 mL of water sample and mixed evenly to obtain spiked sample No. 2. 400 μL of the above Fe 3+ The solution was added to 99.6 mL of water sample and mixed evenly to obtain spiked sample No. 3. 800 μL of the above Fe 3+ The solution was added to 99.2 mL of water sample and mixed evenly to obtain spiked sample No. 4. At this time, the Fe 3+ The spiked concentrations were 10 μM, 20 μM, 40 μM, and 80 μM, respectively.

[0183] (3) The Fe content in acidified water samples and four spiked water samples was detected using the method proposed in this study. 3+ Content. Calculate the spike recovery under the four spike conditions. The calculation method is as follows:

[0184]

[0185] Where R represents the spike recovery, Y0 represents the spike concentration in the acidified water sample, Y1 represents the Fe content detected in the acidified water sample, and Y2 represents the Fe content detected in the spiked water sample. The Fe concentrations and spike recoveries for the five water samples measured by μPADs are shown in Table 3.

[0186] Table 3 Fe in actual water samples and spiked samples based on μPADs 3+ The test results

[0187]

[0188] As can be seen from the table, the spike recovery rate is over 94%. Therefore, this method has high accuracy and good practical detection capability, and can be used for Fe 3+ Quantitative detection of concentration.

[0189] 4. Method Precision

[0190] Precision indicates the reproducibility of measurement and is a prerequisite for ensuring accuracy. To this end, this experimental example examines the precision of this method. The experimental steps are as follows:

[0191] (1) First, μPADs were used to detect Fe at concentrations of 20 μM, 40 μM, and 80 μM. 3+ The solution was measured 10 times for each concentration, and the color images were collected using an industrial camera.

[0192] (2) Then, the color feature values ​​of the color image are extracted by median filtering and image segmentation. The density measurement value corresponding to the color image is calculated.

[0193] (3) Calculate the standard deviation of the values ​​measured in 10 experiments at the same concentration level.

[0194] The standard deviations of the test results at the three concentration levels are shown in Table 4. As can be seen from the table, the relative standard deviation is within 3.0%, and the method precision is good.

[0195] Table 4 Different concentrations of Fe based on μPADs 3+ Relative standard deviation of 10 measurements of the solution

[0196]

[0197] 5. Cost Analysis

[0198] This study analyzed the cost of a single μPAD test based on the amount of reagents and consumables used. The cost of the test can be divided into three components.

[0199] (1) Preparation of blank μPADs. μPADs are mainly composed of a 5 cm × 5 cm hydrophobized cellulose paper sheet and four 16 mm diameter circular cellulose paper sheets. A 20 cm × 20 cm cellulose paper sheet can be used to make eight blank μPADs. The hydrophobization treatment of the paper sheet consumes approximately 0.2 g of white paraffin wax.

[0200] (2) Pretreatment of μPADs. Blank μPADs were treated with two reagent additions. The first treatment required 100 μL of a mixture containing 50 g / L hydroxylamine hydrochloride and 8 g / L 1,10-phenanthroline. The second treatment required 60 μL of a 1.4 g / L polyacrylic acid solution. Both the mixture and the polyacrylic acid solution were prepared using 6.3 M acetate buffer.

[0201] (3) Sample testing. When μPADs are used to detect Fe content in water, the water sample needs to be acidified. The test requires a total of 75 μL of water sample and 25 μL of dilute hydrochloric acid solution with a pH of 1.50, equivalent to a total consumption of 100 μL of dilute hydrochloric acid solution.

[0202] The total usage of each reagent and consumable in the above three parts was counted and the purchase price was combined to analyze the detection cost. The results are shown in Table 5. The single detection cost of the method proposed in the present invention is about 0.70 yuan. 3+ The cost of the concentration method is about 5.0 to 35 yuan per sample. It can be seen that the analysis cost of the method proposed in the present invention is much lower than that of the existing methods.

[0203] Table 5 Single detection cost analysis of μPADs

[0204]

[0205]

[0206] Through the above embodiments and experimental examples, it can be seen that the present invention has constructed a method for detecting the content of Fe in water samples, optimized the color development conditions, image processing methods and the model for calculating the results during the detection process, and achieved high-accuracy detection. The method of the present invention significantly reduces the detection cost while ensuring the performance of anti-interference ability, detection limit, spike recovery rate and precision. In addition, compared with the existing detection of Fe in water, 3+ Compared with the concentration method, the present invention has the unique advantages of simple operation, portability and rapid detection results, and has a good application prospect.

Claims

1. A rapid detection method for iron ions in environmental water, characterized in that: The steps include: Step 1: treating the water sample to be tested with a chemical color development system to cause the chemical color development system to change color, and taking a picture of the chemical color development system; Step 2: Segmenting a region of interest from the image; Step 3, calculating the change value of each component of the RGB color space and the HSI color space in the region of interest; Step 4, using the change values ​​of at least two components obtained in step 3 as features, analyzing the content of iron ions in the water sample to be tested through a multivariate regression equation model; In step 1, the chemical color development system is attached to the microfluidic paper chip, the chemical color development system includes hydroxylamine hydrochloride and 1,10-phenanthroline, the attachment amount of hydroxylamine hydrochloride is 0.622-1.243 mg / cm2, the attachment amount of 1,10-phenanthroline is 0.100-0.199 mg / cm2, and the color development time is 15-25 min; In step 3, the change value of the component is calculated according to the following formula: Where M represents the component of RGB color space or HSI color space, which is one of R, B or S; Mi is the value of M of group i in n parallel experimental groups, is the blank control value of the chemical colorimetric system M with blank water sample added; In step 4, the change value of the component selected as the feature is 、 and , the multivariate regression equation model is: in, is the change value of R, is the change value of B, is the change value of S.

2. The detection method according to claim 1, wherein: In step 2, the region of interest is segmented using edge segmentation.

3. The detection method according to claim 1, wherein: In step 2, the image is subjected to median filtering to segment the region of interest.

4. A rapid detection system for iron ions in environmental water, characterized in that: include: A camera module for taking pictures of the chemical color development system; a segmentation module, for segmenting a region of interest from the image of the chemical color development system; A calculation module, configured to calculate a change value of each component of the RGB color space and the HSI color space in the region of interest; The change values ​​of at least two components are used as features to analyze the content of iron ions in the water sample to be tested through a multivariate regression equation model; The chemical color development system is attached to the microfluidic paper chip, and the chemical color development system includes hydroxylamine hydrochloride and 1,10-phenanthroline. The attachment amount of the hydroxylamine hydrochloride is 0.622-1.243 mg / cm2, the attachment amount of the 1,10-phenanthroline is 0.100-0.199 mg / cm2, and the color development time is 15-25 min. The change value of the component is calculated according to the following formula: Where M represents the component of RGB color space or HSI color space, which is one of R, B or S; Mi is the value of M of group i in n parallel experimental groups, is the blank control value of the chemical colorimetric system M with blank water sample added; The change value of the component selected as the feature is 、 and , the multivariate regression equation model is: in, is the change value of R, is the change value of B, is the change value of S.

5. A computer-readable storage medium, characterized in that: A computer program for implementing the rapid detection method for iron ions in environmental water according to any one of claims 1 to 3 is stored thereon.

Images